Recently, a variety of machine learning techniques have been applied to different healthcare sectors, and the results appear to be promising. One such sector is healthcare planning, in which patient data is used to produce statistical models for predicting the load on different units of the healthcare system. This research introduces an attempt to design and implement a privacy-preserving synthetic data generation method adapted explicitly to patients’ health data and for healthcare planning. A Privacy-preserving Conditional Generative Adversarial Network (PPCGAN) is used to generate synthetic data of Healthcare events, where a well-designed noise is added to the gradients in the training process. The concept of differential privacy is used to ensure that adversaries cannot reveal the exact training samples from the trained model. Notably, the goal is to produce digital patients and model their journey through the healthcare system. / Nyligen har en mängd olika maskininlärningstekniker tillämpats på olika hälso- och sjukvårdssektorer, och resultaten verkar lovande. En sådan sektor är vårdplanering, där patientdata används för att ta fram statistiska modeller för att förutsäga belastningen på olika enheter i sjukvården. Denna forskning introducerar ett försök att utforma och implementera en sekretessbevarande syntetisk datagenereringsmetod som uttryckligen anpassas till patienters hälsodata och för vårdplanering. Ett sekretessbevarande villkorligt generativt kontradiktoriskt nätverk (PPCGAN) används för att generera syntetisk data från hälsovårdshändelser, där ett väl utformat brus läggs till gradienterna i träningsprocessen. Begreppet differentiell integritet används för att säkerställa att motståndare inte kan avslöja de exakta träningsproven från den tränade modellen. Målet är särskilt att producera digitala patienter och modellera deras resa genom sjukvården.

Made available in DSpace on 2016-06-02T19:06:05Z (GMT). No. of bitstreams: 1 5287.pdf: 3924606 bytes, checksum: 935b5a09df26eb1b41df901a189a6e2a (MD5) Previous issue date: 2013-05-29 / Universidade Federal de Sao Carlos / Geographic information systems with the aid of spatial databases store and manage crisp spatial data (or exact spatial data), whose shapes (boundaries) are well defined and have a precise location in space. However, several spatial data do not have precisely known boundaries or have an uncertain location in space, which are called vague spatial data. The boundaries of a given vague spatial data may shrink or extend, therefore, may have a minimum and maximum extension. Clouds of pollution, deforestation, fire outbreaks, route of an airplane, habitats of plants and animals are examples of vague spatial data. In the literature, there are currently vague spatial data models, such as Egg-Yolk, QMM and VASA. However, according to our knowledge, they focus only on the formal aspect of the model definition. Thus, real or synthetic vague spatial data is not available for use. The main objective of this master thesis is the development of algorithms for the generation of synthetic vague spatial data based on the crisp models of spatial data vague Egg-Yolk, QMM and VASA. It was also implemented a tool, called VagueDataGeneration, to assist in the process of generation such data. For both the algorithms and the tool, the user is able to set the properties related to the data type of model, such as size, shape, volume, complexity, location and spatial distribution. By using the proposed algorithms and the VagueDataGeneration tool, researchers can generate large samples of vague spatial data, enabling new research, such as testing indexes for vague spatial data or evaluating query processing over data warehouses that store vague spatial data. The validation of the vague spatial data generation was conducted using a case study with data from vague rural phenomena. / Sistemas de informação geográfica com o auxílio de bancos de dados espaciais armazenam e gerenciam dados espaciais exatos, cujas formas (fronteiras) são bem definidas e que possuem uma localização exata no espaço. Entretanto, vários dados espaciais reais não possuem os seus limites precisamente conhecidos ou possuem uma localização incerta no espaço, os quais são denominados dados espaciais vagos. Os limites de um dado espacial vago podem encolher ou estender, portanto, podem ter uma extensão mínima e máxima. Nuvens de poluição, desmatamentos, focos de incêndios, rota de um avião, habitats de plantas e de animais são exemplos de dados espaciais vagos. Na literatura, atualmente existem modelos de dados espaciais vagos, tais como Egg-Yolk, QMM e VASA. No entanto, segundo o nosso conhecimento, estes enfocam apenas no aspecto formal da definição do modelo. Com isso, dados espaciais vagos reais ou sintéticos não estão disponíveis para uso. O principal objetivo deste trabalho de mestrado consiste no desenvolvimento de algoritmos para a geração de dados espaciais vagos sintéticos baseados nos modelos exatos de dados espaciais vagos Egg-Yolk, QMM e VASA. Também foi implementada uma ferramenta, chamada VagueDataGeneration, para auxiliar no processo de geração desses dados. Nos algoritmos propostos e na ferramenta desenvolvida, o usuário define as propriedades referentes ao tipo de dado de um modelo, tais como tamanho, formato, volume, complexidade, localização e distribuição espacial dos dados espaciais vagos a serem gerados. Por meio do uso dos algoritmos propostos e da ferramenta VagueDataGeneration, os pesquisadores podem gerar grandes amostras de dados espaciais vagos, possibilitando novas pesquisas, como exemplo, testar índices para dados espaciais vagos ou testar técnicas de processamento de consultas em Data Warehouses que armazenam dados espaciais vagos. A validação da geração de dados espaciais vagos foi efetuada usando um estudo de caso com dados de fenômenos rurais vagos.

The ability of human beings to perceive the environment around them with their visual system is called motion perception. This means that the attention of our visual system is primarily focused on those objects that are moving. The property of human motion perception is used in this dissertation to infer human activity from data using artificial neural networks. One of the main aims of this thesis is to discover which modalities, namely RGB images, optical flow and human keypoints, are best suited for HAR in omnidirectional data. Since these modalities are not yet available for omnidirectional cameras, they are synthetically generated and captured with an omnidirectional camera. During data generation, a distinction is made between synthetically generated omnidirectional data and a real omnidirectional dataset that was recorded in a Living Lab at Chemnitz University of Technology and subsequently annotated by hand. The synthetically generated dataset, called OmniFlow, consists of RGB images, optical flow in forward and backward directions, segmentation masks, bounding boxes for the class people, as well as human keypoints. The real-world dataset, OmniLab, contains RGB images from two top-view scenes as well as manually annotated human keypoints and estimated forward optical flow. In this thesis, the generation of the synthetic and real-world datasets is explained. The OmniFlow dataset is generated using the 3D rendering engine Blender, in which a fully configurable 3D indoor environment is created with artificially textured rooms, human activities, objects and different lighting scenarios. A randomly placed virtual camera following the omnidirectional camera model renders the RGB images, all other modalities and 15 predefined activities. The result of modelling the 3D indoor environment is the OmniFlow dataset. Due to the lack of omnidirectional optical flow data, the OmniFlow dataset is validated using Test-Time Augmentation (TTA). Compared to the baseline, which contains Recurrent All-Pairs Field Transforms (RAFT) trained on the FlyingChairs and FlyingThings3D datasets, it was found that only about 1000 images need to be used for fine-tuning to obtain a very low End-point Error (EE). Furthermore, it was shown that the influence of TTA on the test dataset of OmniFlow affects EE by about a factor of three. As a basis for generating artificial keypoints on OmniFlow with action labels, the Carnegie Mellon University motion capture database is used with a large number of sports and household activities as skeletal data defined in the BVH format. From the BVH-skeletal data, the skeletal points of the people performing the activities can be directly derived or extrapolated by projecting these points from the 3D world into an omnidirectional 2D image. The real-world dataset, OmniLab, was recorded in two rooms of the Living Lab with five different people mimicking the 15 actions of OmniFlow. Human keypoint annotations were added manually in two iterations to reduce the error rate of incorrect annotations. The activity-level evaluation was investigated using a TSN and a PoseC3D network. The TSN consists of two CNNs, a spatial component trained on RGB images and a temporal component trained on the dense optical flow fields of OmniFlow. The PoseC3D network, an approach to skeleton-based activity recognition, uses a heatmap stack of keypoints in combination with 3D convolution, making the network more effective at learning spatio-temporal features than methods based on 2D convolution. In the first step, the networks were trained and validated on the synthetically generated dataset OmniFlow. In the second step, the training was performed on OmniFlow and the validation on the real-world dataset OmniLab. For both networks, TSN and PoseC3D, three hyperparameters were varied and the top-1, top-5 and mean accuracy given. First, the learning rate of the stochastic gradient descent (Stochastic Gradient Descent (SGD)) was varied. Secondly, the clip length, which indicates the number of consecutive frames for learning the network, was varied, and thirdly, the spatial resolution of the input data was varied. For the spatial resolution variation, five different image sizes were generated from the original dataset by cropping from the original dataset of OmniFlow and OmniLab. It was found that keypoint-based HAR with PoseC3D performed best compared to human activity classification based on optical flow and RGB images. This means that the top-1 accuracy was 0.3636, the top-5 accuracy was 0.7273 and the mean accuracy was 0.3750, showing that the most appropriate output resolution is 128px × 128px and the clip length is at least 24 consecutive frames. The best results could be achieved with a learning rate of PoseC3D of 10-3. In addition, confusion matrices indicating the class-wise accuracy of the 15 activity classes have been given for the modalities RGB images, optical flow and human keypoints. The confusion matrix for the modality RGB images shows the best classification result of the TSN for the action walk with an accuracy of 1.00, but almost all other actions are also classified as walking in real-world data. The classification of human actions based on optical flow works best on the action sit in chair and stand up with an accuracy of 1.00 and walk with 0.50. Furthermore, it is noticeable that almost all actions are classified as sit in chair and stand up, which indicates that the intra-class variance is low, so that the TSN is not able to distinguish between the selected action classes. Validated on real-world data for the modality keypoint the actions rugpull (1.00) and cleaning windows (0.75) performs best. Therefore, the PoseC3D network on a time-series of human keypoints is less sensitive to variations in the image angle between the synthetic and real-world data than for the modalities RGB images and optical flow. The pipeline for the generation of synthetic data with regard to a more uniform distribution of the motion magnitudes needs to be investigated in future work. Random placement of the person and other objects is not sufficient for a complete coverage of all movement magnitudes. An additional improvement of the synthetic data could be the rotation of the person around their own axis, so that the person moves in a different direction while performing the activity and thus the movement magnitudes contain more variance. Furthermore, the domain transition between synthetic and real-world data should be considered further in terms of viewpoint invariance and augmentation methods. It may be necessary to generate a new synthetic dataset with only top-view data and re-train the TSN and PoseC3D. As an augmentation method, for example, the Fourier Domain Adaption (FDA) could reduce the domain gap between the synthetically generated and the real-world dataset.:1 Introduction 2 Theoretical Background 3 Related Work 4 Omnidirectional Synthetic Human Optical Flow 5 Human Keypoints for Pose in Omnidirectional Images 6 Human Activity Recognition in Indoor Scenarios 7 Conclusion and Future Work A Chapter 4: Flow Dataset Statistics B Chapter 5: 3D Rotation Matrices C Chapter 6: Network Training Parameters

<div> This thesis proposes an artificial neural network (NN) model to predict fuel consumption in heavy vehicles. The model uses predictors derived from vehicle speed, mass, and road grade. These variables are readily available from telematics devices that are becoming an integral part of connected vehicles. The model predictors are aggregated over a fixed distance traveled (i.e., window) instead of fixed time interval. It was found that 1km windows is most appropriate for the vocations studied in this thesis. Two vocations were studied, refuse and delivery trucks.</div><div><br></div><div> The proposed NN model was compared to two traditional models. The first is a parametric model similar to one found in the literature. The second is a linear regression model that uses the same features developed for the NN model.</div><div><br></div><div> The confidence level of the models using these three methods were calculated in order to evaluate the models variances. It was found that the NN models produce lower point-wise error. However, the stability of the models are not as high as regression models. In order to improve the variance of the NN models, an ensemble based on the average of 5-fold models was created. </div><div><br></div><div> Finally, the confidence level of each model is analyzed in order to understand how much error is expected from each model. The mean training error was used to correct the ensemble predictions for five K-Fold models. The ensemble K-fold model predictions are more reliable than the single NN and has lower confidence interval than both the parametric and regression models.</div>